Picture of Neutron Poisons
In nuclear reactor design, we describe the cross-section of different nuclides in a unit called a “barn”. It has units of area. So what does that mean?
Well, take a look at this picture. This shows five important nuclides plotted against each other, with their size determined by their “barns”. You can see that one of them is absolutely HUGE.
That is xenon-135, as far as I know, the nuclide with the largest cross-section to absorb thermal neutrons.
Next on the list of trouble is samarium-149, which is really big, but not nearly as big as xenon-135. Again, to the best of my knowledge, samarium-149 is number #2 on the list of trouble.
For comparison purposes, I show the relative cross-sections of three fissile nuclides, uranium-233, uranium-235, and plutonium-239. These three are fuel in a nuclear reactor, and the bigger these are the better when it comes to making reactors small. By most descriptions of cross-section, U-233, U-235, and Pu-239 have big cross-sections, but you can see that they’re pretty small compared to Xe-135 and Sm-149. Like comparing the inner planets to Jupiter and Uranus.
Why does this matter? Because if we had it our way, we wouldn’t want any Xe-135 or Sm-149 gobbling up neutrons in our reactor. Neutrons they eat are neutrons that we can’t use to make energy by splitting fissile nuclides like Pu-239 or U-233.
Xe-135 has two major differences from Sm-149. The first is that it is radioactive. It goes away if you leave it for awhile. It has a half-life of about nine hours. Sm-149 doesn’t go away. It is not radioactive. It is stable.
The other difference is that xenon is a noble gas, and is easy to remove from a fluid fuel like the salts that we want to use in liquid-fluoride reactors. So getting Xe-135 out of the mix isn’t terribly difficult. Samarium on the other hand is pretty challenging to extract from the salt mixture. It’s one of a family of elements called the “lanthanides“, and they all have very similar chemical properties to each other, because their outermost electron layer (the one that does all the chemical bonding) is the same while they fill up the inner electron layers as you progress up the list of lanthanides. So it’s hard to come up with a chemical process that is particularly good at picking off samarium without picking off all of the other lanthanides at the same time.
I would really love to have someone figure out a nifty way to remove samarium while a fluoride reactor is running. Maybe this guy knows how to do it. I wish I did.
26 thoughts on “Picture of Neutron Poisons”
Xe-135 has such a large neutron cross-section that the vast majority of Xenon-135 atoms will never decay in a reactor – they absorb a neutron and become stable Xe-136 before they get a chance to decay.
One can think of Xe-135 as being equivalent to negative one neutrons, because each atom of Xe-135 is likely to absorb a neutron. It's fission yield is about 6.3% So one way of looking at it is that at equilibrium operation, 6.3% of fission reactions will result in one less neutron.
It's a fairly large hit on neutron economy, but not a deal-breaker. You can still have a reactor with plenty of neutron economy even with Xe-135. The place that it makes the big difference is stability. The fact that Xenon-135 decays and that it absorbs neutrons as it is irradiated means that when a reactor is shut down or brought to low power, it can be very unstable and the characteristics change as Xe-135 decays. For this reason, when a reactor is shut down it may be necessary to wait for the Xe-135 to decay away before restarting.
BUT… there is a cost to removing this stuff. If you leave it in, the vast majority will absorb a neutron and become Xe-136, which is not really a concern of any kind. If removed it will instead decay and it decays to Cs-135.
Removing the Xenon-135 can just about double the yield of Cs-135 from a reactor. Many find that very objectionable because Cs-135 is a long lived radioactive nucleotide.
But Cs-135 is only mildly radioactive, going through a single beta decay to Ba-135 with a half-life of 2.6 million years. Yes some people will worry about it but people will worry about all manner of things that aren't worth worrying about. I think this is one of those cases.
Back-of-envelope estimate for the time available to extract the Sm-149, for anyone else thinking about this:-
A 1000 MW thermal neutron LFTR power plant requires an inventory of about 1000 kg of U-233 to run, and 'burns' about the same quantity per year. An individual U-233 atom in such a reactor therefore has a chance of about 1/365 per day of being hit by a neutron and fissioned.
The thermal neutron fission cross section of U-233 is 530 barns, the absorption cross-section of Sm-149 is about 40,000 barns, about 75 times greater (figures are from JENDL). Therefore the chance of a Sm-149 atom in the above reactor absorbing a neutron is about 75/365 per day or about 1/5. So to halve the loss of neutrons to samarium, you have got to process all the core salt – probably about 20 cubic metres – in five days. Around 170 litres/hr, just under 2 oz/sec
How much is this thing allowed to cost?
What intrigues me is the possibility of some sort of little electrolytic unit, whose electrode would be immersed in the salt and would operate at just the right potential to pick SmF3 (and only SmF3) out of the salt mixture and reduce it on the electrode. Is such a thing possible?
I wonder why you would even want to have a process that picks off samarium but not other lanthanides. Why not just skim all the lanthanides off?
Well, I suppose it would be valuable to take all the lanthanides out, if such a thing could be done relatively simply. Way back when, ORNL looked at overloading the salt with cerium trifluoride, then cooling the salt and causing the lanthanides to preferentially freeze first, and skim them off. Cerium has a much smaller neutron absorption cross section than the other lanthanides like samarium, europium, and neodymium.
I never really liked that approach much because it seemed like you were cramming the salt full of something you'd rather not anyway. If an electrolytic cell could remove lanthanides online, all the better.
Many of the lanthanides reach nuclear stability quite quickly, but they're still significant neutron poisons. Samarium-149 is the best example of this.
If Luke's calculation is correct, and the resource I found putting 136Xe at 2MM barns is also correct, we have about 4 1/2 hours to get the Xenon out in order to halve the neutron loss.
Xenon comes from Iodine, right? So maybe we could also try to get that out before it decays.
I recall some old ORNL research that thought along these lines (removing as iodine rather than xenon). I think they found that it was not very effective, due to the vastly easier step of removing xenon as a gas vs. removing iodine chemically.
I guess my take on it is – how much Sm-149 is actually produced in the reactor on a regular basis, and how hurtful is it to the neutron economy?
If the proportion is small enough – and if said, it becomes stable after one hit of a neutron – then my wild-ass guess is that it would be far more expensive to remove (both in terms of extra equipment and reliability of the reactor) than it would be to just live with it. Or, perhaps it would be useful to apply an extra source of thermal neutrons (getting as cheap as they are).
Sm-149 is already stable. But it still has a huge appetite for neutrons.
Send this challenge directly to nnadir and to Kim Johnson, who both have impressive radiochemistry expertise.
Sm-149 is stable – but is Sm-150 stable? Sm-151? I still think that some quantification is in order; if the problem causes the lifetime of the fuel to go down by a couple of years (ie: 28 instead of 30) then no big deal, if the problem halves the useful lifetime of the fuel, that's a different story..
I think the issue isn't "what has the worst cross-section" but rather "what absorbs the most neutrons in an actual reactor"; you've got to know how much gets made and what the concentration is.
If you had a number for the % of neutrons that get eaten by Sm149 you'd have a more useful indication of how much it matters.
@Steve, you're certainly right about Xe135 in a LWR, where you don't need to spend your neutrons on breeding Pu from U238. On the other hand, a LFTR has to breed (or almost breed) in order to tap Thorium as a fuel, and the breeding margin is pretty thin, compared to Pu-cycle fast reactors. Practically, Th-cycle thermal breeders could have a much smaller fissile inventory, which could make up for this.
[If we put all the commercial Pu we had into one of the best contemporary LMFBR designs, we'd have enough Pu to fuel a breeder fleet about 1/3 the size of our current LWR fleet. The Pu-cycle can access a lot of energy, but the power is limited]
Couldn't you just 'sift' the larger particles out in some way?
I've not thought of any subtle ways to get lanthanides out, they just like being fluorides too much. The brute force method is posted on the forum – link should be my name, above
As far as I am aware the only way to separate this stuff effectively from other lanthanides is with an electrochemical process – basically ion exchange. It's not even a once-through process to refine high quality lanthanides. It has to be a kind of cascade to repeatedly selectively concentrate one type over the others.
It's complex even when it's done for industrial purposes – preparing specialty rare earth materials for the market.
Doing it continuously in a reactor? Seems a little daunting.
I get it's a problem, but it's not a deal killer. Despite it's neutron cross section, it's present in low enough concentrations that a MSR will run quite happily without the need to extract it.
Going to such extremes to pull out all the lanthium seems to be taking perfectionism to the next level.
Long lived fission byproducts are not a big deal to me. I don't worry about Cs-135, but remember that long lived radioactive materials, however low the toxicity, are the driving force behind the hysteria over spent fuel.
Are the engineering benefits of removing xenon worth the political cost? Take a step back and don't look at Cs-135 from a scientific viewpoint but from a PR one. The public has become so condition to react to the prospect of synthetic isotopes in spent fuel with sheer unrestrained terror that I don't think this can be ignored.
what's the density difference? Could it be skimmed effectively off the top or bottom of a slow moving stream kept at constant temp? Or perhaps a helical pipe with the stream at high velocity, like a centrifuge?
I'm no expert, but I know samarium is used in some high-strength magnets. If the flow passes through a magnetic field, maybe the samarium will get pushed to one side.
Is there a chart of the different melting/boiling points for critical substances? Where does Samarium flouride boil versus Thorium Flouride?
Here's a quick Wiki, answers the first part of Ed's June 22 question. High mw SM appears to decay to Europium, which has a much smaller cross section.
From Wikipedia, the free encyclopedia
Jump to: navigation, search
Naturally occurring samarium (Sm) is composed of four stable isotopes, 144Sm, 150Sm, 152Sm and 154Sm, and three extremely long-lived radioisotopes, 147Sm (1.06 × 1011y), 148Sm (7 × 1015y) and 149Sm (>2 × 1015y), with 152Sm being the most abundant (26.75% natural abundance). 146Sm is also fairly long-lived (108y), but occurs naturally as only the tiniest trace remains from its original supernova nucleosynthesis.
151Sm has a half-life of 90 years, and 145Sm has a half-life of 340 days. All of the remaining radioisotopes have half-lives that are less than two days, and the majority of these have half-lives that are less than 48 seconds. This element also has five meta states with the most stable being 141mSm (t½ 22.6 minutes), 143m1Sm (t½ 66 seconds) and 139mSm (t½ 10.7 seconds).
The long lived isotopes,146Sm, 147Sm, and 148Sm primarily decay by alpha decay to isotopes of neodymium. Lighter unstable isotopes of samarium primarily decay by electron capture to isotopes of promethium, while heavier ones decay by beta minus decay to isotopes of europium
Isotopes of samarium are used in samarium-neodymium dating for determining the age relationships of rocks and meteorites.
Samarium-151 is a medium-lived fission product and acts as a neutron poison in the nuclear fuel cycle. The stable fission product Samarium-149 is also a neutron poison.
Where did you find the Xe135 X-section data? I've been looking at the data on the nist and bnl sites and don't see that isotope listed.
I got all the data from the ENDF files in the ORIGEN depletion code.
On electrolysis uranium, thorium, and everything else would come out before samarium except alkali/alkaline earth metals and the lighter lanthanides. If U already removed by fluorination, electrolysis might work for removing lanthanide and transition metal FPs, but would not work for removing cesium and strontium.
On the other hand technetium and other transition metals with volatile high fluorides would come out with uranium on fluorination and not be around for electrolysis. Technetium hexafluoride in particular has a boiling point very close to that of uranium hexafluoride.